Platelet-derived growth factor (PDGF) has been identified as one of the principal mitogens for many types of mesenchymal-derived cells grown in culture (reviewed by Ross et al. Cell 46:155-169, 1986). Due to both its ubiquitous distribution within the body and its chemotactic and mitogenic activities, PDGF has been proposed to play a role in a wide variety of multicellular responses, including wound healing, atherosclerosis, growth and development and neoplasia (Ross et al., ibid , 1986).
PDGF was first identified (Ross et al., Proc. Nat'l. Acad. Sci. USA 71:1207-1210, 1974; Kohler and Lipton, Exp. Cell Res. 87:297-30, 1974) and purified (Antoniades, Proc. Nat'l. Acad. Sci. USA 78:7314-7317, 1981; Deuel et al., J. Biol. Chem. 256:8896-8899, 1981; Heldin et al., Biochemistry J. 193:907-913, 1981; Raines and Ross, J. Biol. Chem. 257:5154-5160, 1982) from human platelets, where it is sequestered in the alpha granules and released upon platelet activation. It has subsequently been isolated from a variety of cell types (for review see Ross et al., ibid , 1986). These cell types include endothelial cells, vascular smooth muscle cells, monocytes and macrophages, derma fibroblasts, several tumor cell lies and a variety of virally transformed cell lines. Platelets have remained a major source of PDGF because outdated platelet material available through blood banks provides a large source of starting material. With the advent of recombinant DNA technology, PDGF has now been expressed in yeast cells (Murray and Kelly, EP 177,957, 1986), thus providing a non-mammalian source of material for therapeutic use.
PDGF is a cationic glycoprotein of approximately 31 kDa which exists as a disulfide-bonded, two-chain molecule. The protein exhibits size heterogeneity, with multiple species ranging from 27 to 31 kDa being observed (Raines and Ross, ibid., 1982). Upon reduction of the molecule, multiple protein species of 14 to 18 kDa are generated (Raines and Ross, ibid., 1982; Kaplan et al., Blood 53:1043-1052, 1979). These reduced species lack the biological activity of the native protein (Raines and Ross, ibid., 1982).
Amino acid sequence analysis of purified PDGF has revealed sequences for two distinct polypeptides, termed A-chain and B-chain, which have been reported to be present in equimolar amounts (Antoniades and Hunkapiller, Science 220:963-965, 1983; Waterfield et al., Nature 304:35-39, 1983; Doolittle et al., Science 221:275-277, 1983; Johnsson et al., EMBO J. 3:921-928, 1984). It is generally believed that the major form of PDGF in platelets is an A-B heterodimer (Johnsson et al., ibid. 1984). Naturally occuring A-chain homodimers and B-chain homodimers have been isolated from osteosarcoma cell-conditioned media (Heldin et al., Nature 319:511-514, 1986) and porcine platelets (Stroobant and Waterfield, EMBO J. 3:2963-2967, 1984), respectively. Kelly et al., (EMBO J. 4:3399-3405, 1985) have reported the production of mitogenically active recombinant B-chain homodimer using transformed yeast cells.
Amino acid sequence analysis of the A- and B-chains isolated from human platelets has shown them to be 54% homologous and to have lengths of 104 and 109 amino acids, respectively (Johnsson et al., ibid., 1984). The B-chain sequence has been shown to be highly homologous to the putative transforming protein (p28-sis) of the simian sarcoma virus (SSV) (Waterfield et al., ibid.; Doolittle et al., ibid.).
Both PDGF A-chain and B-chain cDNAs have recently been obtained (Betsholtz et al., Nature 320:695-699, 1986; Tong et al., Nature 328:619-621, 1987; Collins et al., Nature 328:621-624, 1987), including cDNAs encoding A-chain peptides of 110 and 125 amino acids.
In addition to its mitogenic and chemotactic activities, PDGF has been reported to trigger a wide variety of events following its binding to cell-surface receptors. At least two classes of PDGF receptors have been identified in our laboratory. These have been designated the "B-receptor," which binds only BB-homodimers (BB isoform), and the "A/B-receptor," which binds all three isoforms (AA-homodimers, BB-homodimers- and AB-heterodimers) of PDGF. The existence of such receptor classes may indicate differences in the biological events triggered by the different PDGF isoforms.
Various purification methods have been used to isolate PDGF. These methods generally involve numerous steps and are characterized by low overall yields of purified protein. Isolation of PDGF from human platelet lysates has been described by Heldin et al. (ibid., 1981). This method utilizes sequential chromatography on CM-Sephadex, Blue Sepharose, Bio-Gel P-150 and Sephadex G-200. This method, which provides an overall yield of 6%, was carried out on fresh platelets, a starting material which is not readily available. In another example, Johnsson et al. (Biochem. Biophys. Res. Comm. 104:66-74, 1982) used a series of three sequential chromatographies on CM-Sephadex, Blue Sepharose and Bio-Gel P-150 followed by high pressure liquid chromatography on a Lichrosorb RP 8 column to isolate PDGF derived from platelet lysates. Raines and Ross (ibid., 1982) reported a method to purify PDGF from outdated platelet-rich plasma which resulted in a maximum overall yield of 21%. In this method platelet lysates were subjected to four sequential chromatographies on CM-Sephadex, Sephacryl S-200, Heparin-Sepharose and Phenyl-Sepharose. Antoniades (U.S. Pat. No. 4,479,896, 1983) has reported a method for purifying PDGF from platelet lysates with a 1.5% yield. According to this method, platelet lysates are subjected to sequential precipitations with a primary fractionating reagent, comprising 1M NaCl and 1M acetic acid, and an alcoholic reagent, at a final volumetric concentration of 75%, to extract PDGF polypeptides. The partially purified PDGF polypeptides are then precipitated with acetone. The resultant precipitate is dissolved in chromatographic fluid and subjected to sequential chromatographies on CM-Sephadex, Bio-Gel P-150, and Blue Sepharose. These purification methods are complex and expensive, requiring extensive handling of samples resulting in large losses of material. These problems make the currently available purification methods unsuitable for commercial use.
The previously described methods also provide a PDGF protein product which is heterogeneous and of unknown dimer composition. Commercially available PDGF preparations are also of unknown composition. The dimer composition of the preparation may prove to be important in the therapeutic use of PDGF in view of the discovery of multiple PDGF receptors with different ligand binding specificities. None of the methods currently used to isolate PDGF provides separation of its isoforms (AA, BB and AB) from heterogeneous preparations while retaining the native conformation of the molecules. The only reported method for identification of A-chain and B-chain populations requires the purification of reduced forms of PDGF followed by HPLC analysis (Johnsson et al., ibid., 1982). This method does not lend itself to commercial application because renaturation of the reduced polypeptides does not restore native biological activity.
Previously described antibodies are unsuitable for isolation of specific PDGF isoforms. Available polyclonal antibodies against PDGF (Kelly et al., ibid.) do not demonstrate specificity for the individual isoforms. Monoclonal antibodies which recognize reduced forms of PDGF have been developed using a synthetic peptide derived from the amino-terminal amino acid sequence of the B-chain of PDGF as an immunogen (Niman et al., Proc. Nat'l. Acad. Sci. USA 82:7924-7928, 1985; Niman, Nature 307: 180-183, 1984). However, these antibodies detect the individual chains of PDGF only after reduction of the intact molecule an-d are therefore unsuitable for purifying or detecting the intact, active isoforms of PDGF.
In view of PDGF's clinical applicability in the treatment of injuries in which healing requires the proliferation of fibroblasts or smooth muscle cells and its further value as an important component of a defined medium for the growth of mammalian cells in culture, the production of useful quantities of PDGF is clearly invaluable. With the identification of two receptor classes for PDGF which have different ligand binding properties and with the potential that the individual isoforms of PDGF will stimulate biochemical events unique to each isoform, there is a clear need in the art for methods which allow the isolation of individual isoforms of PDGF in their native, active conformation. Present purification methods are complex and expensive, involving multiple manipulations of the starting material which result in substantial losses of active PDGF. The resultant PDGF product is heterogeneous in nature and contains unknown quantities of the various PDGF isoforms.